...

A high-level assessment of a wide variety of open source solutions was undertaken that could potentially address all or some of the Ed-Fi Level 2 Data Validation System Requirements.  A high-level analysis of the tools that were considered can be found at  Survey of Rule Engines.  However, the final approach taken was to take a deeper look at three different types of solutions from this list. The first one was to look at an open-source business rules engine . The second was to experiment with using a database-centric testing framework. The third was to look at what a pure-SQL implementation without external tool dependencies might look like.

Option 1: Business rule engine solution

Our initial hypothesis when we started this evaluation was that one of the many open source business rules engines would be able to address a significant part of our requirements.  During our analysis we settled in on two Java-based business rules engines, Drools (https://www.drools.org/) and Open Rules (https://openrules.com/).  For our prototype we decided to try Open Rules, which is geared more towards business analysts instead of Java developers.

...

If an organization is already using a business rules infrastructure, it could be adapted to address the validation engine requirements.  This would be especially well suited for complex validations that had multiple decision points. However, for an organization that is just looking to solve for the requirements identified, the overhead and complexity of this type of solution would be hard to justify.

Option 2: Testing framework solution

Test driven and continuous integration development methodologies are widely considered best practices in contemporary software development shops, especially those embracing agile development.  The ubiquity of these approaches has led to a wide variety of automated testing frameworks, some of which are very database-centric. Our hypothesis is that such a tool could be useful for implementing a data validation testing infrastructure.

...

Testing infrastructures do provide a tool set that would allow for a standardized methodology for storing and defining tests.  Potentially this could be leveraged for data validation tests, but the toolset would have to be heavily customized so that it would output specific instances of non-compliant data and so that it would output results to the API.  If an organization was already heavily invested in a testing infrastructure this work could be justified, otherwise it might be a hard sell because the amount of effort to implement the customizations is probably less then just doing something from the ground up (per the previous example).

Option 3: Pure SQL Ed-Fi Data Validation Proof-Of-Concept

Overview

This solution shows a simplified version of a data validation engine built using only dynamic SQL. 

...